[PATCH v3 2/4] pwm: rockchip: Allow polarity invert on rk3288

[Doug Anderson]

Thierry,

On Wed, Aug 20, 2014 at 3:04 AM, Thierry Reding
<thierry.reding at gmail.com> wrote:
> On Tue, Aug 19, 2014 at 09:07:54AM -0700, Doug Anderson wrote:
>> The rk3288 has the ability to invert the polarity of the PWM.  Let's
>> enable that ability.
>>
>> To do this we increase the number of pwm_cells to 3 to allow using the
>> PWM_POLARITY_INVERTED flag.  Since the PWM driver on rk3288 is very
>> new, I thought this was OK.
>
> I don't see any files in arch/arm/boot/dts using either of the
> rockchip,vop-pwm or rockchip,rk3288-pwm compatible strings, so there's
> no reason to consider this stable ABI yet. As far as I'm concerned the
> last sentence can just as well be dropped.
>
> Besides, patches have been posted to support #pwm-cells = <2> and
> #pwm-cells = <3> at the same time which should give you backwards-
> compatibility for free.

Done.  I'm happy that the subsystem is being improved to handle
pwm-cells more dynamically, too!  That's a nice improvement.


> A couple more comments inline.
>
>> diff --git a/drivers/pwm/pwm-rockchip.c b/drivers/pwm/pwm-rockchip.c
> [...]
>> +int rockchip_pwm_set_polarity(struct pwm_chip *chip, struct pwm_device *pwm,
>> +                           enum pwm_polarity polarity)
>
> This should be static.

Done.


>> +{
>> +     struct rockchip_pwm_chip *pc = to_rockchip_pwm_chip(chip);
>> +
>> +     if (!pc->data->has_invert)
>> +             return -ENOSYS;
>> +
>> +     /*
>> +      * No action needed here because pwm->polarity will be set by the core
>> +      * and the core will only change polarity when the PWM is not enabled.
>> +      * We'll handle things in set_enable().
>> +      */
>> +
>> +     return 0;
>> +}
>
> An alternative here would be to provide a separate pwm_ops with
> .set_polarity = NULL for the versions of the IP block that don't support
> polarity inversion yet, but this works for me too.

Good point.  I'll make the change since it paves the way for other
pwm_ops that are different.


>> @@ -173,6 +201,7 @@ static const struct rockchip_pwm_data pwm_data_v2 = {
>>               .ctrl = 0x0c,
>>       },
>>       .prescaler = 1,
>> +     .has_invert = true,
>>       .set_enable = rockchip_pwm_set_enable_v2,
>>  };
>>
>> @@ -184,6 +213,7 @@ static const struct rockchip_pwm_data pwm_data_vop = {
>>               .ctrl = 0x00,
>>       },
>>       .prescaler = 1,
>> +     .has_invert = true,
>>       .set_enable = rockchip_pwm_set_enable_v2,
>>  };
>
> Can you please add a '.has_invert = false,' line to pwm_data_v1? I know
> it's not strictly necessary but I like it when things are explicitly
> stated.

No longer relevant with different pwm_ops.


>> @@ -230,6 +260,10 @@ static int rockchip_pwm_probe(struct platform_device *pdev)
>>       pc->chip.ops = &rockchip_pwm_ops;
>>       pc->chip.base = -1;
>>       pc->chip.npwm = 1;
>> +     if (pc->data->has_invert) {
>
> There should be a blank line between the above two.

Done.